The parietal cortex is critical for spatial attention, and for transferring spatially accurate visual information to the motor system. Humans and monkeys explore space with their eyes, and therefore saccadic eye movements provide an excellent model system for spatial processing in the parietal cortex. A major problem in understanding the genesis of spatially accurate eye movements is to understand how the brain can accurately represent visually guided saccade targets despite a constantly moving eye. The double-step task is a laboratory metaphor for this problem: monkeys and humans can make two sequentially accurate saccades to recently vanished stimuli despite the changing retinal position of the stimuli during the intervening saccade. Humans with right parietal lesions cannot perform this task (Duhamel et al., 1992). Two model theories have been proposed to solve the problem of the double-step task, (1) the gain field theory and (2) the perisaccadic remapping theory. In 1983, Andersen and Mountcastle showed that eye position modulates firing rate of parietal neurons encoding a visual memory guided delayed saccade task in a linear manner, increasing monotonically with orbital eccentricity. A number of neural models proposed that absolute position of a visual target in supraretinal coordinates can easily be calculated from gain fields t solve the double step task (Pouget and Sejnowski; Salinas and Abbott). However the source of eye position modulation of gain fields is not known. Wang et al (2007) discovered an eye position signal in the primary somatosensory cortex (area 3a). Preliminary results show that reversible cooling of area 3a eliminates gain fields. Due to their slow response, we predict the monkey can solve the double step task without gainfields but not visual responses. The second solution to the double step task is the perisaccadic remapping theory which relies on copies of the motor signal (efference copy) to remap the retinal position of an object into the coordinate system. The final stage of this project will be to reversibly inactivate the FEF through cooling. We predict that inactivation of FEF will degrade the monkey's performance in the double step task, will affect remapping of LIP neurons, but will not affect gain fields.